Performance of lymphocyte count by flow cytometry is difficult to evaluate because of the lack of desirable range for precision performance. The aim of this work was to propose recommendations for acceptable precision variability. Data of repeatability and reproducibility of two levels of IQC were collected from sixty-four laboratories. For each cell population, acceptable CV was fixed as the mean CV obtained with robust statistical methods plus 3 standard deviations. Performance limits for reproducibility varied from 2.5% to 9.1%, respectively for CD3+ and NK cells expressed as percentage, and from 9.7% to 14.2% respectively for absolute count of CD3+ and NK cells. Reproducibility data results were obtained from a mean of 21±11 data. Performance limits for reproducibility varied from 2.5% to 9.1% respectively for CD3+ and NK cells expressed as percentage, and from 9.7% to 14.2% respectively for absolute value of CD3+ and NK; 90% of the laboratories had CV inferior to the limit values. Both repeatability and reproducibility values were inversely related to the importance of cell population. These results highlight the accuracy of the method despite the variability of instruments, protocols, IQC and counting systems. It is thus possible to recommend a "state-of-the-art" maximum CV to evaluate and follow over time the performance of the method, and to extrapolate the results to rare cells counting.
Keywords: accreditation; flow cytometry; lymphocyte count; precision; quality control.